- Email: [email protected]
Concerns about the design of clinical trials for spinal muscular atrophy
Neuromuscular Disorders 14 (2004) 456–460 www.elsevier.com/locate/nmd
Review
Concerns about the design of clinical trials for spinal muscular atrophy Thomas O. Crawford* Department of Neurology and Pediatrics, Johns Hopkins University, Jefferson 123, 600 N Wolfe Street, Baltimore, MD 21287, USA Received 9 January 2004; received in revised form 25 February 2004; accepted 1 April 2004
Abstract The distinctive clinical course of SMA, characterized by slowing of the rate of degeneration with the passage of time, presents a special challenge to therapeutic clinical trial planning. Much of the actual functional decline may represent either an inevitable consequence of growth or the result of various secondary complications of weakness, making the study of agents intended to improve the course by increasing the level of SMN protein that much more difficult. Studies intended to demonstrate a slowing of the rate of degeneration, modeled upon clinical trials for ALS, are problematic. In contrast, short-term trials designed to demonstrate improved strength have substantial design advantages, but depend upon the demonstration of salutary effects of increased SMN that are plausible but at present only theoretical. This form of study thus has some potential for type II error, falsely rejecting a useful drug. Despite this limitation, logistic and statistical concerns suggest that the best strategy for evaluating any promising new therapy will be to use first a short-term study. q 2004 Published by Elsevier B.V. Keywords: Spinal muscular atrophy (SMA); SMN; Clinical trial design
Recent advances in the biology of SMN have promoted hope that a compound that substantially increases SMN protein levels by increasing overall SMN2 gene transcription, the proportion of full length SMN2 splicing, or both, will be available for human testing in the near-term. This prospect raises, in turn, a new concern. How best might we demonstrate the benefit of such a compound in patients with SMA? Answering this question requires that we know something about the course of SMA. While much has been learned, an old question of whether SMA is a progressive disorder, and if yes to what degree, remains unanswered. Recent neurophysiologic studies of motor unit numbers have demonstrated convincingly that early in the course of SMA there is a precipitous loss of functioning motor units [1] (Swoboda, K. Personal communication). Yet the clinical course of SMA is unusual for a degenerative disorder: the rate of functional decline clearly slows with the passage of time. Most patients appear to manifest the disorder in two indistinct phases, the first marked by the appearance of new weakness, or manifest departure from a normal pathway of advancing development; and the second a slower declining phase, and for some even a ‘plateau’ phase, during which little or no change may be * Fax: þ 1-410-614-2297. E-mail address: [email protected] (T.O. Crawford). 0960-8966/$ - see front matter q 2004 Published by Elsevier B.V. doi:10.1016/j.nmd.2004.04.004
noted over an extended time [2 – 6]. Patients’ transition takes place from the earlier, more apparent phase of decline to the slower chronic phase at very different ages, depending upon the severity of the disease. At one end of the spectrum are infants with severe SMA who appear to enter the slow chronic phase as early as the first months of life; at the other end are individuals with milder forms of SMA who may not manifest weakness until late teen or even early adult years. Given this unusual clinical course, the ideal, most powerful, clinical trial for an SMN-enhancing agent would evaluate the effect on individuals who are yet to enter into the manifest initial declining phase during which functioning motor units are clearly decreasing. Unfortunately, at present such patients are rarely identified, given the requirement for a sentinel affected sibling. A practical pre-symptomatic screening test of the general population is unlikely to be developed in the near future. Moreover, even if a means for screening populations for homozygous SMN1 deletion were developed, at present we lack the ability from SMN2 copy number alone to estimate beyond a broad range the probable kinetics of motor unit loss, or declining muscle power [7]. Thus this ‘ideal’ clinical trial is not practical in the near term. It appears, then, that clinical trials will first have to target patients in the slower chronic phase of the disorder. Whether
T.O. Crawford / Neuromuscular Disorders 14 (2004) 456–460
this phase is a time when the number of functioning motor units declines at a very slow rate, or instead a time during which there is absolutely no further decline in the number of functioning motor units, remains unknown. Part of the reason for our ignorance about this basic question is that most patients with SMA manifest weakness during the childhood years of growth and development, during which time everything we can assess changes naturally—and thus our baseline for comparison is a moving target. How exactly would a decreased but stable number of functioning motor units manifest during childhood? In addition to this natural complexity, the course of SMA is likely influenced by a number of other confounding variables that influence power and function. I find it useful to consider the clinical manifestation of power and function in children with SMA to be the result of factors that can be grouped into four classes. Class 1 factors are those that directly result from the limited quantity of SMN protein. The number of functioning motor units, which is the premier pathologic feature of SMA, belongs to this class. Class 2 factors are those other genetic or environmental influences that directly affect SMN protein abundance or its role in the motor neuron. Factors of this class affect SMN2 transcription, SMN2 pre-mRNA splicing, SMN protein degradation, or a more downstream process that influences motor neuron susceptibility to diminished SMN protein abundance. Class 3 factors are those that accrue as a natural consequence of motor unit loss, due to any cause, that by nature are beyond the influence of non-specific or SMNspecific treatments. One example of a class 3 variable is the influence of growth on strength and function: do changes in weight and size play any role in declining power? Is there an age-related limit on the capacity for collateral reinnervation by sprouting? A second example involves the natural response of motor neurons to diminished numbers—the post-polio phenomenon. To what extent do overtaxed surviving motor neurons either lose the capacity for maintaining an expanded motor unit territory over time or manifest an increased rate of age-related neuronal dropout? Class 4 factors are the complications of weakness due to any cause. With all neuromuscular disorders, but in particular with SMA where measurements of power appear fairly stable over time, an enhanced vulnerability to complications of weakness is a significant determinant of clinical expression [3]. Potential complications of weakness manifest in almost every organ system. For example, diminished respiratory muscle strength leads to atelectasis and pneumonia, which decreases lung compliance that in turn increases the muscle power necessary to breath. Orthopaedic deformities of the spine and limbs create mechanical impediments to efficient use of residual muscle power. Immobility may lead to regional disuse atrophy, secondary nerve compressions, or disrupt vascular supply. Severely diminished muscle mass creates increased vulnerability to hypoglycemia with catabolic illness, or electrolyte
457
disturbance with gastrointestinal illness. Obesity, cachexia, and other nutritional or hormonal disturbances may follow from severe distortions of muscle mass associated with widespread denervation. Weakness can distort parenting, social relationships, and self-image in ways that are impossibly complex. What class 4 factors all have in common is that foresight and fastidious care may minimize their expression, and hindsight suggests means by which these complications might have been prevented. From the standpoint of experimental design for testing an SMN-enhancing agent, class 2– 4 factors are confounding variables that impair our ability to detect the changes that might follow from enhanced levels of SMN synthesis. Given this multiplicity of factors which influence disease manifestation during the slow chronic phase, how would a trial to test the effect of a putative SMN-enhancing drug be possible? In general, a successful clinical trial anticipates confounding variables by various adjustments in design [8]. Knowledge of the special problems associated with SMA, and preparation for their amelioration as much as possible, will enhance the probability of a successful trial. † Use of a proper control group, likely to experience similar levels of class 2– 4 confounding variables, is a sine qua non for meaningful results. † Increase the size of the experimental and control cohorts to increase the statistical power of the trial. Unfortunately, the relative strength of class 1– 4 variables to influence the clinical course of SMA during the slow chronic phase is unknown. Thus sample size calculations to identify the portion of the course that could be influenced by (partial) restoration of SMN are at best a guess. My personal opinion is that, for most patients, during the chronic phase of SMA the influence of class 3 and 4 factors outweigh class 1 factors that could be influenced by an SMN-enhancing drug. † Identify class 2 factors. When known, controlling for known confounding genetic or environmental influences on the effect of diminished SMN protein will decrease the variability, and thus increase the sensitivity, of a given trial. † Quantify class 3 and 4 complications in an objective manner. Careful clinicians everywhere seek to minimize the effect of class 4 factors as part of routine clinical care. Standardization of clinical care for SMA across the spectrum of disease, and equally important developing a means for reliable measurement of the effect of acquired complications upon function, would aid in trial design. † Increase the homogeneity of the original cohort to increase the statistical power of a given sample size. SMA exists over a spectrum of ages and over a spectrum of severities. An increase in the homogeneity of patients within and between treatment groups realizes a substantial increase in the statistical power to recognize differences with treatment. As a relatively rare disorder,
458
T.O. Crawford / Neuromuscular Disorders 14 (2004) 456–460
however, increasing the homogeneity of patients comes at a practical cost. A smaller ‘window’ of patient eligibility necessitates either increasing the number of trial sites, which inevitably introduces additional complexities of logistics and standardization, or increasing patient travel, which introduces increased problems of recruitment and confounding variables associated with patient fatigue. † Use of surrogate outcome measures less influenced by the confounding factors. Various anatomic, biochemical, or electrophysical surrogate measures of muscle mass might be less influenced by class 4 variables than the strength and functional outcomes that are most meaningful to patients, families and oversight regulatory agencies. For example, DEXA scanning was used as a surrogate outcome for one open label trial [9]. However, no longitudinal studies of surrogate measures have been reported, and thus their natural history, test – retest stability over time, or relationship to power, function and size in a cohort of children with defined levels of weakness due to SMA will need to be established. Creative thinking about new potential surrogate measures of muscle mass, and longitudinal trials of new and established measures across the range of SMA, would be of great potential value. Whether or not a regulatory agency will approve the use of a surrogate measure for drug approval will depend directly upon its previous validation, in the target population, as a measure of important clinical outcomes such as strength, function and lifespan. Plans to use any surrogate outcome for this purpose will need regulatory approval prior to performance of the trial. † Use of direct measures of the intended biologic response to stratify analysis of the primary outcome variable. A direct measure of in vivo full-length SMN transcript, or SMN protein, would be extremely useful to trial design, though insufficient in itself as a proof of clinical benefit. Use of such a measure to stratify changes in strength or function as a function of measured increases in SMN is a very powerful means to analyze clinical data, in that it controls for the portion of variation attributable directly to pharmacokinetics and drug activity [8]. These latter factors can be studied independently. Development of a stable, reliable, practical, and sensitive in vivo measure of full-length SMN transcript or protein abundance in the low-level range seen in clinical SMA is a worthy goal for research in the near term. As noted, the nature of the transition in SMA from early subacute to slow chronic phase is an old question—one that has attracted the attention of clinical investigators for almost as long as the now settled splitter/lumper question about the number of types of SMA. Unlike the nosology question, however, there is little prospect that a definitive answer will arise from further investigations in genetics or animal
modeling. Whatever the absolute answer be, however, the relative answer is clear: changes in functioning motor unit numbers over time during the slow chronic phase will at most be small, and hence a clinical trial intended to test the ability of an SMN-enhancing agent to alter the rate of motor unit loss will necessarily be difficult. These facts might lead to resigned pessimism. There are, however, practical and defensible alternative designs for a slow chronic phase SMA clinical trial. The key is to focus on other potential class I effects. Besides declining motor unit numbers, are there other effects of SMN abundance that might be more easily assessed? In other words, is it possible that restoration of SMN protein abundance might have some other benefit that could be measured over a short term? There are at least some theoretical means by which increases in SMN may have short-term measurable benefit: † An increase in contractile protein within the existing innervated muscle fiber pool. SMN is a ubiquitously expressed protein involved in spliceosomal assembly and recycling, and hence, protein synthesis [10]. While many of the innervated muscle fibers are hypertrophied not all are of similar large caliber. A paucity of functioning spliceosomes might in turn restrict contractile total protein synthesis in a manner that could be rate limiting for use-based fiber hypertrophy. In partial support of this hypothesis, muscle targeted decrease in the abundance of SMN leads to fiber atrophy and fatty replacement similar to that seen in muscular dystrophy [11]. † An increase in the fidelity of neuromuscular transmission. Whether there is some degree of pre-junctional or junctional failure of transmission in the most feeble sprouts of an expanded motor unit has been a subject of speculation, with some supportive evidence [12]. An increase in the SMN abundance may have some influence on the abnormal distal axonal anatomy of SMA [13,14] that may be associated with transmission failure at higher rates of stimulation [15,16]. † An increase in sprouting. It appears that there might be a relationship in the size of surviving functional motor units to SMN abundance, as patients with milder forms of SMA, having more copies of the SMN2 gene [7], tend to larger motor unit size [17]. SMN is abundant in the growth cone of extending axons [18]. SMN abundance thus may have an influence on the capacity for collateral reinnervation following denervation. In patients with intermediate, as opposed to mild, SMA, there appears to be a population of motor units that has not expanded in response to neighboring denervation [19]. Particularly in those individuals most severely affected, whether or not an increase in motor unit size could result from increasing levels of SMN is a reasonable and testable experiment. † Rescue of motor neurons from a non-functional state. Whether or not a failing motor unit exists in a nonfunctional state, perhaps by disconnecting from upper
T.O. Crawford / Neuromuscular Disorders 14 (2004) 456–460
motor neurons in a manner characteristic of that seen in distal-axotomy induced chromatolysis, is unknown. An increase in SMN expression may influence this nonfunctional but viable state [20]. † Promotion of other salutary effects. SMN is ubiquitously expressed. To date, the known deleterious consequences of diminished SMN abundance are restricted to certain populations of neurons. A biochemical abnormality of fatty acid suggesting compromised oxidative pathways of metabolism is present in infants with severe SMA [21]. Though this appears not to lead to any associated symptoms, this or other compromised biochemical pathways or cellular functions outside the nervous system may be symptomatic in a manner that is at present masked by weakness. Testing the response of any one of these class 1 factors to enhanced SMN abundance would be substantially easier than testing a change in the rate of loss of functioning motor units, since the effect would likely be seen in the form of an improvement in a selected outcome measure over a short duration. Such a design has thus characterized all SMA clinical trials to date [22 – 27] (summarized in Ref. [5]). Short duration, smaller size, practical trials using the resources potentially available in one or a few institutions can be designed in a manner that can yield substantial statistical power. Though the advantages of such a trial design are obvious, two cautions are necessary. First, this form of trial design presumes a high potential for type II error, rejecting a useful drug (that could reduce a slow loss of functioning motor units) because it fails to increase power (by inducing other hypothetical other class I mechanisms). Nonetheless, a well performed negative trial would be useful to the design of a follow-up more definitive trial intended to demonstrate an effect on the number of functioning motor units. Given that such a definitive trial would necessarily entail vast patient and financial resources, it would necessarily be postponed until powerful evidence of a motor unit sparing effect in animal models, and human studies of sustainable safe SMN protein induction, are demonstrated. The second caution is that a short-term trial of this nature looks for effects within the same time frame that characterizes the placebo effect. A series of short term trials—either poorly designed and interpreted open-label, small cohort, or single arm studies; or labeled as a preliminary ‘pilot’ study—have promoted, despite concerns about trial design noted above, the benefit of many different therapies [9,24 – 27]. None of these have yet led to a more definitive controlled trial, yet in the real world there is daily evidence that pressure for a ‘cure’ can trump the paucity of supportive evidence. Proper attention to the trial basics including the use of proper controls, blinding, use of wellvalidated outcome measures with a known natural history in the examined population, appropriate preliminary studies in order to calculate the sample sizes necessary to demonstrate
459
or reject a treatment effect decisively, and finally, demonstrated scientific equipoise in the details of trial design and performance, will be essential to move forward.
References [1] Bromberg MB, Swoboda KJ. Motor unit number estimation in infants and children with spinal muscular atrophy. Muscle Nerve 2002;25: 445–7. [2] Dubowitz V. Infantile muscular atrophy: a prospective study with particular reference to a slowly progressive variety. Brain 1964;87: 707–18. [3] Iannaccone ST, Russman BS, Brown RH, Buncher CR, White M, Samaha F. Prospective analysis of strength in spinal muscular atrophy. J Child Neurol 2000;15:97–101. [4] Crawford TO, Pardo CA. The neurobiology of childhood spinal muscular atrophy. Neurobiol Dis 1996;3:97–110. [5] Merlini L, Estournet-Mathiaud B, Iannaccone S, et al. 90th ENMC International Workshop: European spinal muscular atrophy randomised trial (EuroSMART) 9–10 February 2001, Naarden, The Netherlands. Neuromuscul Dis 2002;12:2001– 210. [6] Carter GT, Abresch RT, Fowlere WM, et al. Profiles of neuromuscular diseases. Spinal muscular atrophy. Am J Phys Med Rehabil 1995; 74(Suppl):S150–9. [7] Feldko¨tter M, Schwarzer V, Wirth R, Wienker TF, Wirth B. Quantitative analyses of SMN1 and SMN2 based on real-time lightCycler PCR: fast and highly reliable carrier testing and prediction of severity of spinal muscular atrophy. Am J Hum Genet 2002;70: 358–68. [8] Max MB. Small clinical trials. In: Gallin JI, editor. Principles and practice of clinical research. San Diego: Academic Press; 2002. p. 207 –24. [9] Kinali M, Mercuri E, Main M, et al. Pilot trial of albuterol in spinal muscular atrophy. Neurology 2002;59:609–10. [10] Paushkin S, Gubitz AK, Massenet S, Dreyfuss G. The SMN complex, an assemblyosome of ribonucleoproteins. Curr Opin Cell Biol 2002; 14:305–12. [11] Cifuentes-Diaz C, Frugier T, Tiziano FD, et al. Deletion of murine SMN exon 7 directed to skeletal muscle leads to severe muscular dystrophy. J Cell Biol 2001;152:1107–14. [12] Denys EH, Norris FH. Amyotrophic lateral sclerosis: impairment of neuromuscular transmission. Arch Neurol 1979;36:202 –5. [13] Cifuentes-Diaz C, Nicole S, Velasco ME, et al. Neurofilament accumulation at the motor endplate and lack of axonal sprouting in a spinal muscular atrophy mouse model. Hum Mol Genet 2002;11: 1439– 47. [14] Coe¨rs C, Woolf AL. Pathological anatomy of the intramuscular motor innervation. In: Walton JN, editor. Disorders of voluntary muscle. London: Churchill Livingstone; 1981. p. 238– 60. [15] Rich MM, Waldeck RF, Cork LC, et al. Reduced endplate currents underlie motor unit dysfunction in canine motor neuron disease. J Neurophysiol 2002;88:3293–304. [16] Einarsson G, Grimby G, Stalberg E. Electromyographic and morphological functional compensation in late poliomyelitis. Muscle Nerve 1990;13:165 –71. [17] Hausmanowa-Petrusewicz I. Electrophysiological findings in childhood spinal muscular atrophies. Rev Neurol (Paris) 1988;144: 716–20. [18] Fan L, Simard LR. Survival motor neuron (SMN) protein: role in neurite outgrowth and neuromuscular maturation during neuronal differentiation and development. Hum Mol Genet 2002;11:1605– 14. [19] Galea V, Fehlings D, Kirsch S, McComas A. Depletion and sizes of motor units in spinal muscular atrophy. Muscle Nerve 2001;24: 1168– 72.
460
T.O. Crawford / Neuromuscular Disorders 14 (2004) 456–460
[20] Yamanouchi Y, Yamanouchi H, Becker LE. Synaptic alterations of anterior horn cells in Werdnig-Hoffmann disease. Pediatr Neurol 1996;15:32 –5. [21] Crawford TO, Sladky JT, Hurko O, Besner-Johnston A, Kelley RI. Abnormal fatty acid metabolism in childhood spinal muscular atrophy. Ann Neurol 1999;45:337–43. [22] Merlini L, Solari A, Vita G, et al. Role of gabapentin in spinal muscular atrophy: results of a multicenter, randomized Italian study. J Child Neurol 2003;18:537 –41. [23] Miller RG, Moore DH, Dronsky V, et al. A placebo-controlled trial of gabapentin in spinal muscular atrophy. J Neurol Sci 2001;191: 127–31.
[24] Mercuri E, Bertini E, Messina S, et al. Pilot trial of phenylbutyrate in spinal muscular atrophy. Neuromuscul Dis 2004;14:130–5. [25] Tzeng AC, Cheng J, Fryczynski H, et al. A study of thyrotropin-releasing hormone for the treatment of spinal muscular atrophy: a preliminary report. Am J Phys Med Rehabil 2000;79:435 –40. [26] Takeuchi Y, Miyanomae Y, Komatsu H, et al. Efficacy of thyrotrophin-releasing hormone in the treatment of spinal muscular atrophy. J Child Neurol 1994;9:287– 9. [27] Angelini C, Micaglio GF, Trevisan C. Guanidine hydrochloride in infantile and juvenile spinal muscular atrophy. A double blind controlled study. Acta Neurol (Napoli) 1980;2:460–5.
Review
Concerns about the design of clinical trials for spinal muscular atrophy Thomas O. Crawford* Department of Neurology and Pediatrics, Johns Hopkins University, Jefferson 123, 600 N Wolfe Street, Baltimore, MD 21287, USA Received 9 January 2004; received in revised form 25 February 2004; accepted 1 April 2004
Abstract The distinctive clinical course of SMA, characterized by slowing of the rate of degeneration with the passage of time, presents a special challenge to therapeutic clinical trial planning. Much of the actual functional decline may represent either an inevitable consequence of growth or the result of various secondary complications of weakness, making the study of agents intended to improve the course by increasing the level of SMN protein that much more difficult. Studies intended to demonstrate a slowing of the rate of degeneration, modeled upon clinical trials for ALS, are problematic. In contrast, short-term trials designed to demonstrate improved strength have substantial design advantages, but depend upon the demonstration of salutary effects of increased SMN that are plausible but at present only theoretical. This form of study thus has some potential for type II error, falsely rejecting a useful drug. Despite this limitation, logistic and statistical concerns suggest that the best strategy for evaluating any promising new therapy will be to use first a short-term study. q 2004 Published by Elsevier B.V. Keywords: Spinal muscular atrophy (SMA); SMN; Clinical trial design
Recent advances in the biology of SMN have promoted hope that a compound that substantially increases SMN protein levels by increasing overall SMN2 gene transcription, the proportion of full length SMN2 splicing, or both, will be available for human testing in the near-term. This prospect raises, in turn, a new concern. How best might we demonstrate the benefit of such a compound in patients with SMA? Answering this question requires that we know something about the course of SMA. While much has been learned, an old question of whether SMA is a progressive disorder, and if yes to what degree, remains unanswered. Recent neurophysiologic studies of motor unit numbers have demonstrated convincingly that early in the course of SMA there is a precipitous loss of functioning motor units [1] (Swoboda, K. Personal communication). Yet the clinical course of SMA is unusual for a degenerative disorder: the rate of functional decline clearly slows with the passage of time. Most patients appear to manifest the disorder in two indistinct phases, the first marked by the appearance of new weakness, or manifest departure from a normal pathway of advancing development; and the second a slower declining phase, and for some even a ‘plateau’ phase, during which little or no change may be * Fax: þ 1-410-614-2297. E-mail address: [email protected] (T.O. Crawford). 0960-8966/$ - see front matter q 2004 Published by Elsevier B.V. doi:10.1016/j.nmd.2004.04.004
noted over an extended time [2 – 6]. Patients’ transition takes place from the earlier, more apparent phase of decline to the slower chronic phase at very different ages, depending upon the severity of the disease. At one end of the spectrum are infants with severe SMA who appear to enter the slow chronic phase as early as the first months of life; at the other end are individuals with milder forms of SMA who may not manifest weakness until late teen or even early adult years. Given this unusual clinical course, the ideal, most powerful, clinical trial for an SMN-enhancing agent would evaluate the effect on individuals who are yet to enter into the manifest initial declining phase during which functioning motor units are clearly decreasing. Unfortunately, at present such patients are rarely identified, given the requirement for a sentinel affected sibling. A practical pre-symptomatic screening test of the general population is unlikely to be developed in the near future. Moreover, even if a means for screening populations for homozygous SMN1 deletion were developed, at present we lack the ability from SMN2 copy number alone to estimate beyond a broad range the probable kinetics of motor unit loss, or declining muscle power [7]. Thus this ‘ideal’ clinical trial is not practical in the near term. It appears, then, that clinical trials will first have to target patients in the slower chronic phase of the disorder. Whether
T.O. Crawford / Neuromuscular Disorders 14 (2004) 456–460
this phase is a time when the number of functioning motor units declines at a very slow rate, or instead a time during which there is absolutely no further decline in the number of functioning motor units, remains unknown. Part of the reason for our ignorance about this basic question is that most patients with SMA manifest weakness during the childhood years of growth and development, during which time everything we can assess changes naturally—and thus our baseline for comparison is a moving target. How exactly would a decreased but stable number of functioning motor units manifest during childhood? In addition to this natural complexity, the course of SMA is likely influenced by a number of other confounding variables that influence power and function. I find it useful to consider the clinical manifestation of power and function in children with SMA to be the result of factors that can be grouped into four classes. Class 1 factors are those that directly result from the limited quantity of SMN protein. The number of functioning motor units, which is the premier pathologic feature of SMA, belongs to this class. Class 2 factors are those other genetic or environmental influences that directly affect SMN protein abundance or its role in the motor neuron. Factors of this class affect SMN2 transcription, SMN2 pre-mRNA splicing, SMN protein degradation, or a more downstream process that influences motor neuron susceptibility to diminished SMN protein abundance. Class 3 factors are those that accrue as a natural consequence of motor unit loss, due to any cause, that by nature are beyond the influence of non-specific or SMNspecific treatments. One example of a class 3 variable is the influence of growth on strength and function: do changes in weight and size play any role in declining power? Is there an age-related limit on the capacity for collateral reinnervation by sprouting? A second example involves the natural response of motor neurons to diminished numbers—the post-polio phenomenon. To what extent do overtaxed surviving motor neurons either lose the capacity for maintaining an expanded motor unit territory over time or manifest an increased rate of age-related neuronal dropout? Class 4 factors are the complications of weakness due to any cause. With all neuromuscular disorders, but in particular with SMA where measurements of power appear fairly stable over time, an enhanced vulnerability to complications of weakness is a significant determinant of clinical expression [3]. Potential complications of weakness manifest in almost every organ system. For example, diminished respiratory muscle strength leads to atelectasis and pneumonia, which decreases lung compliance that in turn increases the muscle power necessary to breath. Orthopaedic deformities of the spine and limbs create mechanical impediments to efficient use of residual muscle power. Immobility may lead to regional disuse atrophy, secondary nerve compressions, or disrupt vascular supply. Severely diminished muscle mass creates increased vulnerability to hypoglycemia with catabolic illness, or electrolyte
457
disturbance with gastrointestinal illness. Obesity, cachexia, and other nutritional or hormonal disturbances may follow from severe distortions of muscle mass associated with widespread denervation. Weakness can distort parenting, social relationships, and self-image in ways that are impossibly complex. What class 4 factors all have in common is that foresight and fastidious care may minimize their expression, and hindsight suggests means by which these complications might have been prevented. From the standpoint of experimental design for testing an SMN-enhancing agent, class 2– 4 factors are confounding variables that impair our ability to detect the changes that might follow from enhanced levels of SMN synthesis. Given this multiplicity of factors which influence disease manifestation during the slow chronic phase, how would a trial to test the effect of a putative SMN-enhancing drug be possible? In general, a successful clinical trial anticipates confounding variables by various adjustments in design [8]. Knowledge of the special problems associated with SMA, and preparation for their amelioration as much as possible, will enhance the probability of a successful trial. † Use of a proper control group, likely to experience similar levels of class 2– 4 confounding variables, is a sine qua non for meaningful results. † Increase the size of the experimental and control cohorts to increase the statistical power of the trial. Unfortunately, the relative strength of class 1– 4 variables to influence the clinical course of SMA during the slow chronic phase is unknown. Thus sample size calculations to identify the portion of the course that could be influenced by (partial) restoration of SMN are at best a guess. My personal opinion is that, for most patients, during the chronic phase of SMA the influence of class 3 and 4 factors outweigh class 1 factors that could be influenced by an SMN-enhancing drug. † Identify class 2 factors. When known, controlling for known confounding genetic or environmental influences on the effect of diminished SMN protein will decrease the variability, and thus increase the sensitivity, of a given trial. † Quantify class 3 and 4 complications in an objective manner. Careful clinicians everywhere seek to minimize the effect of class 4 factors as part of routine clinical care. Standardization of clinical care for SMA across the spectrum of disease, and equally important developing a means for reliable measurement of the effect of acquired complications upon function, would aid in trial design. † Increase the homogeneity of the original cohort to increase the statistical power of a given sample size. SMA exists over a spectrum of ages and over a spectrum of severities. An increase in the homogeneity of patients within and between treatment groups realizes a substantial increase in the statistical power to recognize differences with treatment. As a relatively rare disorder,
458
T.O. Crawford / Neuromuscular Disorders 14 (2004) 456–460
however, increasing the homogeneity of patients comes at a practical cost. A smaller ‘window’ of patient eligibility necessitates either increasing the number of trial sites, which inevitably introduces additional complexities of logistics and standardization, or increasing patient travel, which introduces increased problems of recruitment and confounding variables associated with patient fatigue. † Use of surrogate outcome measures less influenced by the confounding factors. Various anatomic, biochemical, or electrophysical surrogate measures of muscle mass might be less influenced by class 4 variables than the strength and functional outcomes that are most meaningful to patients, families and oversight regulatory agencies. For example, DEXA scanning was used as a surrogate outcome for one open label trial [9]. However, no longitudinal studies of surrogate measures have been reported, and thus their natural history, test – retest stability over time, or relationship to power, function and size in a cohort of children with defined levels of weakness due to SMA will need to be established. Creative thinking about new potential surrogate measures of muscle mass, and longitudinal trials of new and established measures across the range of SMA, would be of great potential value. Whether or not a regulatory agency will approve the use of a surrogate measure for drug approval will depend directly upon its previous validation, in the target population, as a measure of important clinical outcomes such as strength, function and lifespan. Plans to use any surrogate outcome for this purpose will need regulatory approval prior to performance of the trial. † Use of direct measures of the intended biologic response to stratify analysis of the primary outcome variable. A direct measure of in vivo full-length SMN transcript, or SMN protein, would be extremely useful to trial design, though insufficient in itself as a proof of clinical benefit. Use of such a measure to stratify changes in strength or function as a function of measured increases in SMN is a very powerful means to analyze clinical data, in that it controls for the portion of variation attributable directly to pharmacokinetics and drug activity [8]. These latter factors can be studied independently. Development of a stable, reliable, practical, and sensitive in vivo measure of full-length SMN transcript or protein abundance in the low-level range seen in clinical SMA is a worthy goal for research in the near term. As noted, the nature of the transition in SMA from early subacute to slow chronic phase is an old question—one that has attracted the attention of clinical investigators for almost as long as the now settled splitter/lumper question about the number of types of SMA. Unlike the nosology question, however, there is little prospect that a definitive answer will arise from further investigations in genetics or animal
modeling. Whatever the absolute answer be, however, the relative answer is clear: changes in functioning motor unit numbers over time during the slow chronic phase will at most be small, and hence a clinical trial intended to test the ability of an SMN-enhancing agent to alter the rate of motor unit loss will necessarily be difficult. These facts might lead to resigned pessimism. There are, however, practical and defensible alternative designs for a slow chronic phase SMA clinical trial. The key is to focus on other potential class I effects. Besides declining motor unit numbers, are there other effects of SMN abundance that might be more easily assessed? In other words, is it possible that restoration of SMN protein abundance might have some other benefit that could be measured over a short term? There are at least some theoretical means by which increases in SMN may have short-term measurable benefit: † An increase in contractile protein within the existing innervated muscle fiber pool. SMN is a ubiquitously expressed protein involved in spliceosomal assembly and recycling, and hence, protein synthesis [10]. While many of the innervated muscle fibers are hypertrophied not all are of similar large caliber. A paucity of functioning spliceosomes might in turn restrict contractile total protein synthesis in a manner that could be rate limiting for use-based fiber hypertrophy. In partial support of this hypothesis, muscle targeted decrease in the abundance of SMN leads to fiber atrophy and fatty replacement similar to that seen in muscular dystrophy [11]. † An increase in the fidelity of neuromuscular transmission. Whether there is some degree of pre-junctional or junctional failure of transmission in the most feeble sprouts of an expanded motor unit has been a subject of speculation, with some supportive evidence [12]. An increase in the SMN abundance may have some influence on the abnormal distal axonal anatomy of SMA [13,14] that may be associated with transmission failure at higher rates of stimulation [15,16]. † An increase in sprouting. It appears that there might be a relationship in the size of surviving functional motor units to SMN abundance, as patients with milder forms of SMA, having more copies of the SMN2 gene [7], tend to larger motor unit size [17]. SMN is abundant in the growth cone of extending axons [18]. SMN abundance thus may have an influence on the capacity for collateral reinnervation following denervation. In patients with intermediate, as opposed to mild, SMA, there appears to be a population of motor units that has not expanded in response to neighboring denervation [19]. Particularly in those individuals most severely affected, whether or not an increase in motor unit size could result from increasing levels of SMN is a reasonable and testable experiment. † Rescue of motor neurons from a non-functional state. Whether or not a failing motor unit exists in a nonfunctional state, perhaps by disconnecting from upper
T.O. Crawford / Neuromuscular Disorders 14 (2004) 456–460
motor neurons in a manner characteristic of that seen in distal-axotomy induced chromatolysis, is unknown. An increase in SMN expression may influence this nonfunctional but viable state [20]. † Promotion of other salutary effects. SMN is ubiquitously expressed. To date, the known deleterious consequences of diminished SMN abundance are restricted to certain populations of neurons. A biochemical abnormality of fatty acid suggesting compromised oxidative pathways of metabolism is present in infants with severe SMA [21]. Though this appears not to lead to any associated symptoms, this or other compromised biochemical pathways or cellular functions outside the nervous system may be symptomatic in a manner that is at present masked by weakness. Testing the response of any one of these class 1 factors to enhanced SMN abundance would be substantially easier than testing a change in the rate of loss of functioning motor units, since the effect would likely be seen in the form of an improvement in a selected outcome measure over a short duration. Such a design has thus characterized all SMA clinical trials to date [22 – 27] (summarized in Ref. [5]). Short duration, smaller size, practical trials using the resources potentially available in one or a few institutions can be designed in a manner that can yield substantial statistical power. Though the advantages of such a trial design are obvious, two cautions are necessary. First, this form of trial design presumes a high potential for type II error, rejecting a useful drug (that could reduce a slow loss of functioning motor units) because it fails to increase power (by inducing other hypothetical other class I mechanisms). Nonetheless, a well performed negative trial would be useful to the design of a follow-up more definitive trial intended to demonstrate an effect on the number of functioning motor units. Given that such a definitive trial would necessarily entail vast patient and financial resources, it would necessarily be postponed until powerful evidence of a motor unit sparing effect in animal models, and human studies of sustainable safe SMN protein induction, are demonstrated. The second caution is that a short-term trial of this nature looks for effects within the same time frame that characterizes the placebo effect. A series of short term trials—either poorly designed and interpreted open-label, small cohort, or single arm studies; or labeled as a preliminary ‘pilot’ study—have promoted, despite concerns about trial design noted above, the benefit of many different therapies [9,24 – 27]. None of these have yet led to a more definitive controlled trial, yet in the real world there is daily evidence that pressure for a ‘cure’ can trump the paucity of supportive evidence. Proper attention to the trial basics including the use of proper controls, blinding, use of wellvalidated outcome measures with a known natural history in the examined population, appropriate preliminary studies in order to calculate the sample sizes necessary to demonstrate
459
or reject a treatment effect decisively, and finally, demonstrated scientific equipoise in the details of trial design and performance, will be essential to move forward.
References [1] Bromberg MB, Swoboda KJ. Motor unit number estimation in infants and children with spinal muscular atrophy. Muscle Nerve 2002;25: 445–7. [2] Dubowitz V. Infantile muscular atrophy: a prospective study with particular reference to a slowly progressive variety. Brain 1964;87: 707–18. [3] Iannaccone ST, Russman BS, Brown RH, Buncher CR, White M, Samaha F. Prospective analysis of strength in spinal muscular atrophy. J Child Neurol 2000;15:97–101. [4] Crawford TO, Pardo CA. The neurobiology of childhood spinal muscular atrophy. Neurobiol Dis 1996;3:97–110. [5] Merlini L, Estournet-Mathiaud B, Iannaccone S, et al. 90th ENMC International Workshop: European spinal muscular atrophy randomised trial (EuroSMART) 9–10 February 2001, Naarden, The Netherlands. Neuromuscul Dis 2002;12:2001– 210. [6] Carter GT, Abresch RT, Fowlere WM, et al. Profiles of neuromuscular diseases. Spinal muscular atrophy. Am J Phys Med Rehabil 1995; 74(Suppl):S150–9. [7] Feldko¨tter M, Schwarzer V, Wirth R, Wienker TF, Wirth B. Quantitative analyses of SMN1 and SMN2 based on real-time lightCycler PCR: fast and highly reliable carrier testing and prediction of severity of spinal muscular atrophy. Am J Hum Genet 2002;70: 358–68. [8] Max MB. Small clinical trials. In: Gallin JI, editor. Principles and practice of clinical research. San Diego: Academic Press; 2002. p. 207 –24. [9] Kinali M, Mercuri E, Main M, et al. Pilot trial of albuterol in spinal muscular atrophy. Neurology 2002;59:609–10. [10] Paushkin S, Gubitz AK, Massenet S, Dreyfuss G. The SMN complex, an assemblyosome of ribonucleoproteins. Curr Opin Cell Biol 2002; 14:305–12. [11] Cifuentes-Diaz C, Frugier T, Tiziano FD, et al. Deletion of murine SMN exon 7 directed to skeletal muscle leads to severe muscular dystrophy. J Cell Biol 2001;152:1107–14. [12] Denys EH, Norris FH. Amyotrophic lateral sclerosis: impairment of neuromuscular transmission. Arch Neurol 1979;36:202 –5. [13] Cifuentes-Diaz C, Nicole S, Velasco ME, et al. Neurofilament accumulation at the motor endplate and lack of axonal sprouting in a spinal muscular atrophy mouse model. Hum Mol Genet 2002;11: 1439– 47. [14] Coe¨rs C, Woolf AL. Pathological anatomy of the intramuscular motor innervation. In: Walton JN, editor. Disorders of voluntary muscle. London: Churchill Livingstone; 1981. p. 238– 60. [15] Rich MM, Waldeck RF, Cork LC, et al. Reduced endplate currents underlie motor unit dysfunction in canine motor neuron disease. J Neurophysiol 2002;88:3293–304. [16] Einarsson G, Grimby G, Stalberg E. Electromyographic and morphological functional compensation in late poliomyelitis. Muscle Nerve 1990;13:165 –71. [17] Hausmanowa-Petrusewicz I. Electrophysiological findings in childhood spinal muscular atrophies. Rev Neurol (Paris) 1988;144: 716–20. [18] Fan L, Simard LR. Survival motor neuron (SMN) protein: role in neurite outgrowth and neuromuscular maturation during neuronal differentiation and development. Hum Mol Genet 2002;11:1605– 14. [19] Galea V, Fehlings D, Kirsch S, McComas A. Depletion and sizes of motor units in spinal muscular atrophy. Muscle Nerve 2001;24: 1168– 72.
460
T.O. Crawford / Neuromuscular Disorders 14 (2004) 456–460
[20] Yamanouchi Y, Yamanouchi H, Becker LE. Synaptic alterations of anterior horn cells in Werdnig-Hoffmann disease. Pediatr Neurol 1996;15:32 –5. [21] Crawford TO, Sladky JT, Hurko O, Besner-Johnston A, Kelley RI. Abnormal fatty acid metabolism in childhood spinal muscular atrophy. Ann Neurol 1999;45:337–43. [22] Merlini L, Solari A, Vita G, et al. Role of gabapentin in spinal muscular atrophy: results of a multicenter, randomized Italian study. J Child Neurol 2003;18:537 –41. [23] Miller RG, Moore DH, Dronsky V, et al. A placebo-controlled trial of gabapentin in spinal muscular atrophy. J Neurol Sci 2001;191: 127–31.
[24] Mercuri E, Bertini E, Messina S, et al. Pilot trial of phenylbutyrate in spinal muscular atrophy. Neuromuscul Dis 2004;14:130–5. [25] Tzeng AC, Cheng J, Fryczynski H, et al. A study of thyrotropin-releasing hormone for the treatment of spinal muscular atrophy: a preliminary report. Am J Phys Med Rehabil 2000;79:435 –40. [26] Takeuchi Y, Miyanomae Y, Komatsu H, et al. Efficacy of thyrotrophin-releasing hormone in the treatment of spinal muscular atrophy. J Child Neurol 1994;9:287– 9. [27] Angelini C, Micaglio GF, Trevisan C. Guanidine hydrochloride in infantile and juvenile spinal muscular atrophy. A double blind controlled study. Acta Neurol (Napoli) 1980;2:460–5.